The VERSES Research Lab is an international centre for research in statistical physics, biological physics, and complexity science, as well as artifical and human intelligence. Robust work on algorithms, techniques, and foundations for optimisation and learning theory in statistics also happens in the lab. Via our research interests in the free energy principle and its corollaries—Bayesian mechanics, active inference, and predictive coding—the lab is producing highly original work in these areas. VERSES and the Spatial Web Foundation have generously funded these endeavours. Our position on the importance of new, neuroscience- and physics-inspired machine learning theory has recently been articulated in the paper Designing Ecosystems of Intelligence from First Principles.

The research group consists of the following people:

Team

Mahault Albarracin (website)

Christopher L Buckley (website)

Karl J Friston (website)

Conor Heins (website)

Alex B Kiefer (website)

Brennan Klein (website)

Magnus Koudahl (website)

Beren Millidge (website)

Riddhi Jain Pitliya (website)

Maxwell J D Ramstead (website)

Dalton A R Sakthivadivel (you are here)

Toby St Clere Smithe (website)

Safae Essafi Tremblay (website)

Simon C Tremblay (website)

Alexander Tschantz (website)

As well as former team members.
Former members of the VERSES Lab include: Jonas Mago (website), Gabriel Axel Montes (website), Jurgis Pasukonis (website)

Additionally, the lab has strong connections to the EASy group at the University of Sussex, the Theoretical Neurobiology group at University College London’s Wellcome Centre, the Levin lab at Tufts University, and the MSP group at the University of Strathclyde. The lab has a physical base of operations in Los Angeles, California. (Current as of November 2022.)

 

Publications and Preprints

Selected publications are listed below. For full lists, see our members’ webpages, or the lab’s Google Scholar page. Except for where PDFs are provided, publications are open access and can be found at the links indicated.

Mathematics and Physics

Machine Learning, Control Theory, Optimisation

The Free Energy Principle

Neuroscience, Philosophy, Human Intelligence and Perception

Comments on “How Particular is the Physics of the Free Energy Principle?”

 

Mathematics and Physics

A Worked Example of the Bayesian Mechanics of Classical Objects. Dalton A R Sakthivadivel. June 2022. Preprint arXiv:2206.12996. Link. To appear in The Third International Workshop on Active Inference.

On Bayesian Mechanics: A Physics of and by Beliefs. Maxwell J D Ramstead*, Dalton A R Sakthivadivel*, Conor Heins, Magnus Koudahl, Beren Millidge, Lancelot Da Costa, Brennan Klein, Karl J Friston. May 2022. Preprint arXiv:2205.11543. Link.
* equal contributions; listed alphabetically by surname

Towards a Geometry and Analysis for Bayesian Mechanics. Dalton A R Sakthivadivel. April 2022. Preprint arXiv:2204.11900. Link.

 

Machine Learning, Control Theory, Optimisation

Capsule Networks as Generative Models. Alex B Kiefer, Beren Millidge, Alexander Tschantz, Christopher L Buckley. September 2022. Preprint arXiv:2209.02567. Link. To appear in The Third International Workshop on Active Inference.

Successor Representation Active Inference. Beren Millidge, Christopher L Buckley. July 2022. Preprint arXiv:2207.09897. Link. To appear in The Third International Workshop on Active Inference.

pymdp: A Python Library for Active Inference in Discrete State Spaces. Conor Heins, Beren Millidge, Daphne Demekas, Brennan Klein, Karl J Friston, Iain Couzin, Alexander Tschantz. Journal of Open Source Software. May 2022. Link. See arXiv:2201.03904 for technical appendices.

Hybrid Predictive Coding: Inferring, Fast and Slow. Alexander Tschantz*, Beren Millidge*, Anil K Seth, Christopher L Buckley. April 2022. Preprint arXiv:2204.02169. Link.
* equal contributions; listed alphabetically by first name

 

The Free Energy Principle

Path Integrals, Particular Kinds, and Strange Things. Karl J Friston, Lancelot Da Costa, Dalton A R Sakthivadivel, Conor Heins, Grigorios A Pavliotis, Maxwell J D Ramstead, Thomas Parr. October 2022. Preprint arXiv:2210.12761. Link.

On the Map-Territory Fallacy Fallacy. Maxwell J D Ramstead, Dalton A R Sakthivadivel, Karl J Friston. August 2022. Preprint arXiv:2208.04275. Link.

Extended Plastic Inevitable. Maxwell J D Ramstead and Karl J Friston. Constructivist Foundations. July 2022. Link.

Spin Glass Systems as Collective Active Inference. Conor Heins, Brennan Klein, Daphne Demekas, Miguel Aguilera, Christopher L Buckley. July 2022. Preprint arXiv:2207.06970. Link. To appear in The Third International Workshop on Active Inference.

Situated Models and the Modeler: A Comment on “The Markov Blanket Trick: On the Scope of the Free Energy Principle and Active Inference” by Raja, Valluri, Baggs, Chemero and Anderson. Mahault Albarracin and Riddhi J Pitliya. July 2022. Journal link.

Epistemic Communities under Active Inference. Mahault Albarracin, Daphne Demekas, Maxwell J D Ramstead, Conor Heins. Entropy. March 2022. Link.

 

Neuroscience, Philosophy, Human Intelligence and Perception

Mapping Husserlian phenomenology onto active inference. Mahault Albarracin, Riddhi J Pitliya, Maxwell J D Ramstead*, Jeffrey Yoshimi*. August 2022. Preprint arXiv:2208.09058. Link. To appear in The Third International Workshop on Active Inference.
* equal contributions; listed alphabetically by surname

Commentary: The Nature of Beliefs and Believing. Mahault Albarracin and Riddhi J Pitliya. Frontiers in Psychology. July 2022. Link.

From Generative Models to Generative Passages: A Computational Approach to (Neuro) Phenomenology. Maxwell J D Ramstead, Anil K Seth, Casper Hesp, Lars Sandved‑Smith, Jonas Mago, Michael Lifshitz, Giuseppe Pagnoni, Ryan Smith, Guillaume Dumas, Antoine Lutz, Karl J Friston, Axel Constant. Review of Philosophy and Psychology. March 2022. Link.

Active Inference Models do not Contradict Folk Psychology. Ryan Smith, Maxwell J D Ramstead, Alex Kiefer. Synthese. March 2022. Link.

 

Comments on “How Particular is the Physics of the Free Energy Principle?”

A group effort amongst lab members to take stock of the state of the art of the FEP around April and May of 2022. The target article appeared in volume 40 of Physics of Life Reviews (link), whilst the comments, and the authors’ response to those comments, are to appear in volume 41 and volume 42.

 
Particular Flows and Attracting Sets. Conor Heins. June 2022. Preprint. Journal link.

Sparse Coupling and Markov Blankets. Conor Heins and Lancelot Da Costa. June 2022. Preprint. Journal link.

Regarding Flows Under the Free Energy Principle. Dalton A R Sakthivadivel. May 2022. Preprint. Journal link.

Some Minimal Notes on Notation and Minima. Maxwell J D Ramstead and Dalton A R Sakthivadivel. May 2022. Preprint. Journal link.